Large-Scale Production and Concentration of Infectious Epstein-Barr Virus

Transcription

1 /78/ $02.00/0 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 1978, p Copyright C 1978 American Society for Microbiology Vol. 35, No. 1 Printed in U.S.A. Large-Scale Production and Concentration of Infectious Epstein-Barr Virus FREDERICK KLEIN,* JOHN F. ROSENSTEEL, RAYMOND M. HUMMER, ELIZABETH A. HILLMAN,t CHARLES W. RIGGS, AND LEIGH J. CHARMELLA Frederick Cancer Research Center, Frederick, Maryland Received for publication 17 May 1977 Infectious Epstein-Barr virus (EBV) was produced from suspension cultures of P3HR-1 cells. A protocol for the scaled-up production and concentration of the virus was developed. Virus from the culture fluid was concentrated by continuous-flow pelletization or continuous flow with banding in sucrose. EBV prepared by pelletization yielded 1.7 x 104 infectious units/rnl (10Ox concentration) and <3.4 x 107 EVB particles/ml (1,OOOx concentration), whereas EBV prepared by banding yielded 4.6 x 103 infectious units/ml (10Ox concentration) and 1.3 x 10i EBV particles/ml (1,000X concentration). The majority of the virus particles observed were "empty" membrane-associated particles. A statistical test of the correlation between infectivity and virus particle count was made. Epstein-Barr virus (EBV), a member of the MATERIALS AND METHODS herpes group of viruses, remains high on the list as a potential cancer-causing virus in Cell cultures. The EBV humans. producer cell line, P3HR- 1, a subline of P3J (Jijoye) (8), used for the production Originally the agent was discovered in lymphoblastoid cell cultures from Burkitt's lymphoma (Department of Microbiology, Mayo Clinic, Roches- of EBV and Raji cells, was furnished by Gary Pearson (4); however, EBV has been shown to have ter, Minn.). The P3HR-1 cultures contained 10% immunofluorescent cells for virus capsid antigen (VCA) world-wide distribution and is implicated as the causative agent in infectious mononucleosis, and produced 10' infectious units/ml (determined by Burkitt's lymphoma, and nasopharyngeal carcinoma. Recent studies by Miller et al. (14), Klein P3HR-1 and Raji master stock cultures were frozen Raji superinfection assay on 10Ox concentrates). et al. (11), and Menezes et al. (13) suggest that and maintained at liquid nitrogen temperature (gas phase). different strains of EBV exist. It is Working stocks of P3HR-1 and Raji cells were important prepared by concentration of the master stock 10-fold to determine if infectious mononucleosis, Burkitt's lymphoma, and nasopharyngeal carcinoma One-milliliter samples of the concentrate were frozen in growth medium containing 10% dimethyl sulfoxide. are specifically induced by one of the biologically at a controlled rate of 1 C/min and stored in liquid distinct EBV strains. Studies involving the biological, biochemical, immunological, and onco- of working stock was used as described in Table 1. nitrogen. In each production scale-up, one ampoule genic characteristics require large quantities of P3HR-1 and Raji cell cultures were grown in RPMI infectious virus. Only a few reports (9, 18) on 1640 medium supplemented with L-glutamine (Grand large-scale production of EBV from extracellular Island Biological Co.), 10% heat-inactivated fetal bovine serum, 100 U of penicillin per ml, and 100,g of fluid are available, and these studies utilized streptomycin per ml. total virus particle counts (VPC) and not virus Raji cells were grown in 250 ml of medium in 500- infectivity to monitor results. ml serum bottles on a gyratory shaker at 125 to 130 This report presents data on the production rpm and 35 C. Culture medium was replaced daily of EBV from extracellular P3HR-1 fluids in by centrifuging the cultures at 500 x g for 10 min, which infectivity as well as VPC were analyzed. decanting the supernatant, and resuspending cells in The investigation included: (i) P3HR-1 cell fresh medium. Raji cultures were maintained at 2 x growth during inoculum scale-up; (ii) examination of different protocols for concentration of mined by the trypan-blue exclusion technique. P3HR- 106 to 4 x 106 cells/ml and were 95% viable, as detervirus; (iii) characterization of virus concentrates 1 and Raji cultures, as well as virus concentrates, were by electron microscopy, and infectivity; and (iv) routinely monitored for mycoplasma with specific culture media. examination of the relationship between infectivity and particle morphology of EBV. culture fluids were clarified by continuous-flow cen- EBV clarification and concentration. P3HR-1 t Present address: Department of Pathology, University of trifugation in a Sorvall SS-4 or RC2B centrifuge Maryland, School of Medicine, Baltimore, MD equipped with the Szent-Gyorgi and Blum system. 172

2 VOL. 35, 1978 Clarification was done at 12,000 x g at a flow rate of 190 ml/min. The clarified supernatant fluid was subsequently concentrated by either continuous-flow pelletization (CFP) or continuous flow with banding (CFB). In the CFP method, the culture fluids containing EBV were first filtered through 1.2- and then 0.8-aum filters. After filtration, the culture fluid containing EBV was pelletized by the Szent-Gyorgi and Blum continuous-flow centrifugation at 35,000 x g with a flow rate of 100 ml/min. All virus pellets were collected and resuspended in growth medium at 0.01 or of the original volume, dispensed into 1-ml samples, and stored at -70 C. In the CFB method, culture fluids containing EBV were concentrated and partially purified by CFB centrifuge in the RK zonal ultracentrifuge with the RK-3 core (10, 18). Immunofluorescence assay. Infectious EBV was determined by superinfection of Raji cells (6) diluted to 5 x 105 cells/ml in T30 flasks containing 5 ml of RPMI % fetal bovine serum. The cultures were inoculated with appropriate 10-fold dilutions of virus and incubated at 35 C for 4 days. The superinfected cells were removed from the culture medium by centrifugation (200 x g for 10 min), washed with phosphate-buffered saline, spotted on Teflon-coated slides, air dried, and acetone fixed. Human EBV-positive and -negative sera were then applied, fluoresceinlabeled Hyland goat anti-human immunoglobulin was added, and the cells were examined for early antigen by immunofluorescence. The EBV-positive sera were high in VCA and low in early antigen activity as determined by testing with P3HR1 and superinfected Raji cells, respectively. The titer was recorded as the reciprocal of the highest virus dilution giving >0.5% of the cells with positive immunofluorescence when examined for early antigens. Electron microscopy assay. Virus concentrates were pelleted at 35,000 x g and resuspended in distilled water. Virus was visualized by electron microscopy using the negative stain (NS) technique (15). In addition to determining the VPC, the morphological types of herpes particles present in a number of the concentrates were tabulated. The critical-point-drying technique (3) was also used to examine the virus concentrates. RESULTS P3HR-1 cell growth during inoculum scale-up and virus production. The scale-up procedure of P3HR-1 cultures from 1-ml ampoules (working stock) to 6-liter, screw-capped Erlenmeyer flasks is presented in Table 1. Final production volumes of 36 liters consisted of 12 6-liter Erlenmeyer flasks, each containing 3 liters of P3HR-1 culture fluid. This scale-up procedure consisted of 10 steps requiring a total of 32 days. The incubation temperature was 35 C, flasks were shaken by hand once daily, and the culture medium-air volume ratio was 1:5 for maximum cell growth (19), with the exception of the 6- liter flask where it was 1:2 (step 10, Table 1). With a decreased culture-air space ratio in ves- EBV PRODUCTION 173 TABLE 1. Procedure for scale-up of P3HR-1 inocula for production of infectious EBV to 36-liter volume Vol of in- Step oculum (nil) 1 1 ml of frozen working stock added to 24 ml of medium in 125-nil flaska ml of inoculum diluted with 25 ml of growth medium in 250-ml flask ml of inoculum diluted with 50 ml of growth medium in 500-ml flask ml of inoculum diluted with 100 ml of growth medium in 1,000-ml flaskb ml of inoculum diluted to 600 ml with growth medium and distributed in 200-ml volumes in 3 1-liter flasks Each 200 ml of inoculum diluted to 600 ml with growth medium and transferred to 3 3-liter flasks.1, ml of inoculum combined and diluted to 2,400 ml with growth medium and distributed in 200-ml volumes to 12 1-liter flasks.2,400 8 Each 200 ml of inoculum diluted to 600 ml with growth medium and transferred to 12 3-liter flasks.7,200 9 Each 600 ml of inoculum diluted to 1,200 ml with growth medium and transferred to 12 6-liter flasks. 14, Each 1,200 ml of inoculum diluted to 3,000 ml with growth medium and incubated in 6-liter flasks until harvested. 36,000 a Cultural conditions: Vessel, screw-capped Erlenmeyer flasks, incubated statically except for agitation once every day at 350C; viable cell count was monitored and at cell density of 1 x 106 to 2 x 106 viable cells/ml (3 to 5 days) cultures were diluted to -5 x 106 viable cells/ml; culture-air space ratio 1:5, except in step 10. b Sample of culture removed and tested for VCA, virus infectivity, and presence of mycoplasma. If VCA was <10% and infectivity <103 infectious units/ml or cultures were positive for mycoplasma, the inoculum was discarded and a frozen working stock was initiated. sels other than 7-liter flasks, the P3HR-1 cultures did not survive. The kinetics of cell growth and virus replication in the P3HR-1 culture was established using different sized Erlenmeyer flasks to permit effective scale-up of cells and subsequent maxiimal production of EBV. Growth curves of P3HR-1 cell cultures during culture scale-up (200 ml/i-liter flask) (step 4, Table 1) and aging in the 6-liter production flask (3,000 mi/6-liter flask) (step 10, Table 1) are illustrated in Fig. 1. The percentage of VCA-positive cells noted as cells infected for both growth conditions remained at 10% throughout the duration of the experiment. Infectivity levels increased early in the growth cycle in the 1-liter flask and

3 174 KLEIN ET AL. APPL. ENVIRON. MICROB1OL. I0E 0 x c 0.0 co Post - Inoculation, days FIG. 1. Growth curves of P3HR-1 cultures during EBV production in 1- and 6-liter Erlenmeyer flasks. Cultures were sampled at indicated times (data points) and cell viability, percentage of VCA-positive cells, and virus infectivity were determined. remained fairly constant throughout the plateau and declining phases of cell growth. Routinely, when cultures reached 1 x 106 to 2 x 106 viable cells/ml, usually at 3 to 5 days post-transfer, they were diluted with fresh media to 5 x 105 viable cells/ml. With less viable cells, the scaleup time would increase. When cell viability fell below 50%, usually at 7 to 10 days, cultures were harvested for virus. Inoculum scale-up was shortened if cultures at step 4 (Table 1) were maintained as working stock and used for subsequent scale-up. In the 6-liter production flasks essentially no cell growth occurred, but the percentage of VCA-positive cells and infectivity remained constant. Culture fluids harvested from the 6-liter cultures after 6 days yielded virus with the least cellular debris, so that subsequent filtration and clarification was simplified. No increase in infectivity was found with longer aging times. EBV concentration. A typical ultraviolet absorbance profile obtained by CFB is shown in Fig. 2. Fractions 3 through 15 in the 55 to 25% (wt/vol) sucrose regions (1.257 to g/cm3) were pelleted by direct centrifugation in the type 35 rotor at 46,000 x g for 90 min. The distribution of infectivity throughout the gradient is shown in Fig. 3. The gradient was divided into two parts, one containing fractions 5 to 9 and the other containing fractions 10 to 15. Using critical-point drying technique, VPC of 3.7 x 107 and 4.4 x 107/ml were determined, ~I Model RK Ultracentrifuge 254nm - 60 K? GradL Step -- 10,a,.II...nJnr Fraction (50 ml) FIG. 2. Typical ultraviolet absorbance profile (254 nm) from CFB (model RK ultracentrifuge), using a sucrose gradient of culture fluid containing EB V FRACTIONS 50 ml FIG. 3. Ultraviolet absorbance profile (254 nm) from CFB (model RK ultracentrifuge), using a sucrose gradient of culture fluid containing EBV and infectivity titers of individual fractions. respectively. When the NS technique was used, VPC was obtained only on the entire gradient and not on the two individual samples. Thirteen different production lots of P3HR-1 culture were used to compare the yields of EBV ae

4 VOL. 35, 1978 using the CFP and CFB methods (Table 2). Several of these lots were subdivided before concentration, so that 17 concentration experiments were involved in the comparison. From each of these scale-up production lots, a 50-ml sample was taken before the concentration process, and infectivity was determined. The average infectivity titers are given in Table 2. VPC were not obtainable on these preconcentration samples since the actual numbers of virus particles were insufficient for detection by electron microscopy. Differences between these mean "before concentration" titers were not statistically significant when compared with the variation among lots. The infectivity titers of the concentrates were compared by analysis of variance, which indicated that the mean titers of these two concentration methods were not significantly different. However, VPC with the CFP method were below detectable levels with the NS method, whereas with the CFB method virus particles were routinely observed and counted. Samples of CFP concentrates were therefore pelleted in BEEM (Better Equipment Electron Microscopy, Bronx, N.Y.) capsules, embedded, and thin sectioned. Primarily, mature herpestype virions trapped in excessive cellular debris were observed. Other manipulations such as polyethylene glycol precipitation (1), double banding in sucrose (10), and the use of different gradient materials in conjunction with pelletization and banding were evaluated and gave comparable results (Table 2). The cell pellets were sonically treated to release intracellular or adsorbed virus, and the amount of infectious virus found was insignificant when compared with the infectivity found in the extracellular fluids. Electron microscopic evaluation. Electron EBV PRODUCTION 175 microscopy was used to examine the concentrates obtained by the different methods and to determine whether infectivity correlated with the number of virus particles present in the concentrates. An attempt was also made to determine whether morphology of the EBV was influenced by the concentration procedure. VPC were routinely possible when CFB and double banding methods were used (Table 2), and the mean counts per milliliter of concentrate were 1.3 x 10" and 2.1 x 108, respectively. In the CFP method, EBV particles were observed only on five out of eight lots. Regardless of concentration method, primarily single virus particles were observed (Fig. 4a, c, and e). More than half (54%) of all herpes particles counted, using the NS technique, were "empty" membrane-associated particles (Fig. 4b) (5). The second most frequently observed particles (26%) were naked and "empty" (Fig. 4a). "Full" particles containing nucleic acid (Fig. 4b and e) and enveloped particles (Fig. 4c) were rarely observed. The integrity of the envelope as well as the size and shape varied greatly. Using the critical-point drying method, the capsomeres and nucleocapsid structure were well preserved. However, a defined unit membrane envelope was not seen. Instead, an amorphous, fuzzy coat was attached to most particles. The particle size varied with the method used to prepare the specimen. Larger particles with a diameter of 95 to 106 nm were found with the NS technique, whereas with the critical-point drying technique, particles with a diameter of 76 to 82 nm were found. The nucleoid of the herpes particle prepared by critical-point drying (Fig. 4e, arrow, 50 nm) was very similar to that observed in thin sections of P3HR-1 cells (Fig. 4f, arrow, 46 nm). Virus particles were not readily demonstrable in thin-sec- TABLE 2. Comparison of methods for concentration of EBV Avg infectivity levels/ml Mean Method Method~ ~ No. ~ of expt ~ Av Vol cultuer (loox) VPCb/ml b(/mox (liters) Before concn After concn After concn CFP x x 104 -c CFB, with K-3 core x x x i08 PEG followed by pelletization x x 105 PEG followed by single banding X 1i0 5.0 x x 107 PEG followed by double banding x x x 104 Double banding x x x 104 CFB (potassium tartrate) x 104 pld 1.6 x x 107 p2d 2.7 x x 107 CFB, with K-6 core x x x 107 a Raji cell superinfection assay using early antigen as indicator of virus infectivity. b PC using NS method. c -, EBV particles were not routinely observed. d Two peaks were obtained from 10 and 50% equal volume (wt/vol) potassium tartrate gradient; the first peak was in the 29 to 32% region, and the second was in the 22 to 26% region.

5 176 KLEIN ET AL. APPL. ENVIRON. MICROBIOL. FIG. 4. Electron micrographs of EBV concentrated from P3HR-1 cells. (A-C) 2% phosphotungstic acid, ph 4.2, negative stain; (D and E) critical-point dried and stained with 1% uranyl acetate; (F) thin section of pelleted concentrate stained with uranyl magnesium acetate and lead citrate. (A-E) x150,000; (F) x75,000.

6 VOL. 35, 1978 tioned P3HR-1 cells which were 10% VCA positive. However, if herpesvirus particles were observed, they were usually immature and trapped in cellular debris. Correlation coefficients were calculated to estimate the degree of relationship between infectivity and each of the four particle types as demonstrated by NS. In no case did the correlation of infectivity with the count of any particle type attain statistical significance (P = <0.05). DISCUSSION P3HR-1 cultures can be scaled-up to large volume using Erlenmeyer flasks without loss of cell viability and proportion of VCA-positive cells. The cultural conditions include static incubation at 35 C, except for agitation once every day and a culture-air space ratio of 1:5. Under these conditions, cultures reached 1 x 106 to 2 x 106 viable cells/ml at 3 to 5 days. At this time cultures contained the highest viability and infectivity titers. Since recovery from frozen ampoules of cells with high viability and high proportions of VCA is variable, maintenance of cultures in 1-liter flasks is advantageous. Virus infectivity appears to be independent of cell viability and growth. Cultures in which the percentage of VCA-positive cells decreased to 5% or less yielded less virus. Virus infectivity increases early in the growth cycle and remains constant through the plateau and decline of growth (Fig. 1). Since these are chronically infected cells with only a small percentage of cells capable of synthesizing virus, and the factor(s) that converts a VCA-negative cell to a VCApositive cell is unknown, no direct relationship between cell growth and virus synthesis can be made. Several methods for the concentration of EBV were compared. Those methods using banding procedures consistently yielded higher particle counts, whereas the infectivity titers were comparable to, but slightly lower than, titers of pelletized concentrates. Pelletized virus concentrates yielded VPCs just within the limits of detection. Various manipulations in conjunction with the banding method represented by only two or three experiments (Table 2) yielded virus concentrates with similar infectivity titers and particle counts. Although the concentration techniques are standard, it was necessary to apply them to infectious EBV production and document the results. A direct relationship between VCA-positive cells and presence of virus particles, as suggested by Hinuma et al. (7), was not established. Electron microscopy did not readily demonstrate virus particles. When virus was observed, pri- EBV PRODUCTION 177 marily immature herpes particles trapped in cellular debris were seen. These observations are similar to those observed by other investigators using P3HR-1 cell lines (12). Correlation coefficients were computed for the concentration methods in which virus particles were counted, and in no case did the degree of correlation between infectivity and observed particles attain statistical significance. The lack of correlation between infectivity and particle count could be attributable to the volume of culture fluid being processed, the concentration factor, and the presence of cellular debris. Cellular debris interfered with the counting accuracy, since trapped virus particles are often difficult to visualize. In addition, its presence could affect the spread and adherence of material to the microscope grid. Likewise, no correlation between particle morphology and infectivity was determined, as was demonstrated in the herpes simplex system (16). VPC-to-infectious virus ratios obtained from banding methods (Table 2) were approximately 3,000:1. Since VPC were either barely detectable or below 3.4 x 107/ml for the pelletization method, no such ratio can be established for these methods. Data accumulated for herpes simplex virus give a ratio of 10:1 (2). Thus, these data would indicate a need for a more permissive cell system for EBV synthesis, although the P3HR-1 culture can be used and the virus synthesized can be concentrated by a variety of methods. ACKNOWLFDGMEbNIS This research was sponsored by the Public Health Service National Cancer Institute (NCI) under contract NOI-CO , with Litton Bionetics, Inc. We acknowledge J. Elser, D. Warfield, and R. Bradley for electron microscopy support, B. Brown for technical assistance in conducting infectivity assays, L Thibault in the statistical analysis, and G. P. Shibley and the FCRC Viral Resources Laboratory for performing the second banding manipulation. Appreciation is also expressed to G. Pearson (NCI) and D. L. Fine (Frederick Cancer Research Center [FCRC]) for discussions in coordination of this program, and J. Geisberg and A. Huter for secretarial assistance. The FCRC Product Evaluation Laboratory is also acknowledged for the mycoplasma testing. LITERATURE CITED 1. Adams, A Concentration of Epstein-Barr virus from cell culture fluids with polyethylene glycol. J. Gen. Virol. 20: David, B. D., R. Dulbecco, H. N. Eisen, H. S. Ginsberg, and W. B. Wood Microbiology, p Heber Medical Division, Harper and Row Publisher, New York. 3. DeHarven, E., D. Beja, D. P. Evenson, S. Bass, and G. Schidlovsky Structure of critical point dried oncornaviruses. Virology 56: Epstein, M. A., B. G. Achong, and X. M. Barr Virus particles in culture lymphoblasts from Burkitt's

7 178 KLEIN ET AL. lymphoma. Lancet i: Heine, U Intranuclear viruses, p In H. Busch (ed.), The cell nucleus, vol. 3. Academic Press Inc., New York. 6. Henle, W., G. Henle, B. A. Zajac, G. Pearson, R. Waubke, and M. Scriba Differential reactivity of human serums with early antigens induced by Epstein-Barr virus. Science 169: Hinuma, Y., M. Konn, J. Yamuguchi, and J. T. Grace, Jr Replication of herpes-type virus in a Burkitt lymphoma cell line. J. Virol. 1: Hinuma, Y., M. Konn, J. Yamaguchi, D. J. Wudurski, J. R. Blakeslee, and J. T. Grace Immunofluorescence and herpes-type virus particles in the P3HR- 1 Burkitt lymphoma cell line. J. Virol. 1: Jensen, E. M., F. T. Buscheck, and D. Riccardo Large-scale production of Burkitt lymphoma cells and EB virus in tissue culture. Gann Monogr. 10: Johnson, R. W., A. Perry, 0. R. Robinson, Jr., and G. P. Shibley Method for reproducible largevolume production and purification of Rauscher murine leukemia virus. Appl. Environ. Microbiol. 31: Klein, G., B. Sugden, W. Teebold, and J. Menezes Infection of EBV-genome-negative and -positive human lymphoblastoid cell lines with biologically different preparations of EBV. Intervirology 3: Maurer, B. A., T. Imamura, and S. N. Wilbert Incidence of EB virus-containing cells in primary and secondary clones of several Burkitt lymphoma cell lines. Cancer Res. 30: APPL. ENVIRON. MICROBIOL. 13. Menezes, J., W. Leibold, and G. Klein Biological differences between Epstein-Barr virus (EBV) strains with regard to lymphocyte transforming ability superinfection and antigen induction. Exp. Cell Res. 92: Miller, G., J. Robinson, L. Heston, and M. Kipman Differences between laboratory strains of Epstein- Barr virus based on immortalization, abortive infection and interference. Proc. Natl. Acad. Sci. U.S.A. 71: Monroe, J. H., and P. M. Brandt Rapid semiquantitative method for screening large numbers of virus samples by negative staining electron microscopy. Appl. Microbiol. 20: Spring, S. B., and B. Roizman Herpes simplex virus products in productive and abortive infection. I. Stabilization with formaldehyde and preliminary analyses by isopycnic centrifugation in CsCl. J. Virol. 1: Tolpin, I., R. Boyden, A. DePadova, P. Brandt, and P. Sottong The zonal centrifuge applied to the purification of a low-yield virus in human leukemia cells. Biotechnol. Bioeng. 10: Tolpin, I., and P. Sottong Large-volume purification of tumor viruses by use of zonal centrifuges. Appl. Microbiol. 23: Weirether, F. J., J. S. Walker, and R. E. lincoln A precise method for replicating suspension cultures of mammalian cells. Appl. Microbiol. 16: Downloaded from on March 7, 2019 by guest